Layer 16 · Neuron 7090 · suppressed

L16N7090-1

Word quirks:

“ originally”“ initially”“ original”“ mainly”“ primarily”“ originals”“ initial”“ primary”

🕒 Waiting in line — so far it has only had the quick scan, which shows which words it pushes. It was not selected in this round.

Identity (according to the specimen itself)

I am Qwen3.5, developed by Alibaba Cloud. My primary mission is to assist users in solving problems across a wide range of scenarios, from business applications to personal use cases. I can help you with:

Answering questions: Providing accurate information on topics like technology, science, and general knowledge.
Content creation: Generating reports, articles, or creative works.
Code generation: Assisting with programming tasks, including debugging and optimization.
Multilingual support: Communicating fluently in over 100 languages.
Integration: Seamlessly integrating into

Nearest neighbors

Specimens with similar replies or word quirks. Similarity runs from −1 to 1; 1 is identical. These are leads, not shared mechanisms. How profiles are built.

By word quirks Profile recipe

Specimen	Similarity	Word quirks
1. L18N7485-1💤	0.86	“ primary”“ originals”“primary”
2. L18N5253+1🕒	0.85	“ originally”“ original”“ initial”
3. L14N5343+1🕒	0.85	“ originally”“ initially”“ first”
4. L5N5957+1🕒	0.85	“ initial”“ original”“ originally”
5. L16N5967-1💤	0.84	“ initially”“ originally”“ initial”
6. L13N7018-1🕒	0.84	“ originally”“ initially”“ earlier”
7. L17N6501-1🕒	0.84	“ originally”“ initially”“ Originally”
8. L18N5083+1💤	0.83	“ earliest”“ initially”“最早”
9. L18N7698+1💤	0.83	“ initial”“ creation”“ final”
10. L11N6986+1🕒	0.83	“ originally”“ original”“ Originally”

Shadow specimen, apart from the ten: L16N7090+1🕒, -0.04.

Cousins · specimens sharing a neuron

No specimen with replies shares this neuron yet.

Technical details

Quick scan: measure prediction shifts on fixed reference text. Short test: generate replies to a small set of prompts. Longer test: extend to a larger set. Full test: run the full prompt suite. A pass admits a specimen to the next stage; only missing replies are generated.

Layer 16, neuron 7090, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.65, top-1 agreement 0.73) is outside the band that feeds the short test (KL 1 to 2.5, agreement 0.5 to 0.65), so it stays at the quick scan unless someone queues it by hand. A hand pick goes straight to the full test. See Method for the tiers.

Quick scan · prediction measurements

Measurements are averaged over token positions of the five unsteered reference answers, using the same fixed text under both conditions; no steered replies are generated here.

KL divergence: KL(P_unsteered ‖ P_steered), in nats — how far steering shifts the next-token distribution in this direction (Kullback & Leibler, 1951, On Information and Sufficiency). Unsteered ≈ 0.001.
Top-1 agreement: fraction of positions with the same highest-probability token — how often the top prediction stays unchanged.
ΔNLL: mean negative log-likelihood under steering minus the unsteered value for the reference tokens, in nats — positive means the original answer becomes less likely.
Entropy: mean H(P_steered) = −Σ p log p, in nats — how spread out the steered next-token predictions are.

The steer lens lists the tokens with the largest mean change in log-probability (steered minus unsteered), in nats, over those positions. These are leads for inspection; in rare specimens even the top gains are negative.

Setting	KL divergence	Top-1 agreement	ΔNLL	Entropy	Steer lens
L16N7090+1🕒 shadow specimen t0-v0.1	0.799	0.65	0.993	1.41	“ shouldn”“...\”“ crucial”“ further”“ choice”“ necessary”“ __”“:"”
L16N7090-1 t0-v0.1	0.654	0.73	0.881	1.20	Top boosted tokens ↑

Method

A specimen is a fixed steering configuration of Qwen3.5-4B: one MLP neuron or a pair, each with a signed dose. Unsteered is the control. L identifies the layer; N, the neuron.

Additive activation steering. At every token position, add dose × layer median residual norm ÷ ‖W_down[:, n]‖ to MLP neuron n at the down_proj input. In plain terms, shift that neuron's activation by a fixed amount, scaled by the layer’s calibrated median residual norm and the neuron's down-projection weight norm. Positive doses push up; negative doses push down. Dose is normalized intervention size, not a percentage or probability.